app.py

import os
from huggingface_hub import InferenceClient

client = InferenceClient(api_key=os.environ["HF_API_TOKEN"])

response = client.text_generation(
    model="microsoft/Phi-4-mini-flash-reasoning",
    prompt="User: What is the capital of France?\nAssistant:",
    max_new_tokens=50,
    temperature=0.7
)

print(response)